Stm32F302C8t6 IRQ shenanigan and D+ pull up resistor

[cdg66]

I've made a board with an stm32F302C8 and I've notice that the usb IRQ is weird the irq is somewhat bound to the can IRQ and separated in two LP and HP.
My question is where should I call the tud_int_handler();
I'm currently calling in both LowPriority and HighPriority IRQ but it dont show up on my computer. Should i call it in USBWakeUp_RMP_IRQn() instead cause ive noticed this in dcd_stm32_fsdev.c?

// Some STM32F302/F303 devices allow to remap the USB interrupt vectors from
// shared USB/CAN IRQs to separate CAN and USB IRQs.
// This dynamically checks if this remap is active to disable the right IRQs.
#ifdef SYSCFG_CFGR1_USB_IT_RMP
if (SYSCFG->CFGR1 & SYSCFG_CFGR1_USB_IT_RMP)
{
NVIC_DisableIRQ(USB_HP_IRQn);
NVIC_DisableIRQ(USB_LP_IRQn);
NVIC_DisableIRQ(USBWakeUp_RMP_IRQn);
}
else
#endif
{
NVIC_DisableIRQ(USB_HP_CAN_TX_IRQn);
NVIC_DisableIRQ(USB_LP_CAN_RX0_IRQn);
NVIC_DisableIRQ(USBWakeUp_IRQn);
}

Thx for your help

[pigrew]

I suggest calling tud_int_handler from all three interrupt handlers. However, based on reading the reference manual, I think it's only actually required to be called from the LP handler (since I believe that all events trigger the LP interrupt, with some also triggering the other ones).

(I've never tried remapping the interrupts, so the below is what I understand based on the reference manual, and I have not tried to have different handlers for different interrupts.)

It doesn't hurt (except for wasting CPU time) to call the interrupt handler too often. The handler checks the interrupt flags, and will do nothing in the case that there is no event flag set.

The manual for your MCU ( RM0365 Rev 8) seems to have an error, it lists the remap bit (SYSCFG_CFGR1.5), but later says it is reserved, so I don't know if it works in the F302C8 or not.

My guess is that there is some other issue causing your enumeration failure.

[hathach]

Have you tried to use the stock example first ? It is always good to start with something working first

[cdg66]

you mean the stm32 middleweare?

[hathach]

No I mean this repo example. You should try it out first, the build instruction can be found in https://github.com/hathach/tinyusb/blob/master/docs/getting_started.md

[cdg66]

What ive tried is STM32 Middleweare, UAC2 Headset example and HID Composite. In all case I'm able to get an interupt for a connection and unpulg of the usb and nothing else. What else could be causing that i dont have no more packet. I'm using stm32IDE and cube mx and the config like in the discussion #633. I've made it working with an stm32F4 discovery board but not on the F3

[hathach]

I don't use stm32 cube and middleware, and couldn't help any further. keep debugging, I am sure it is something minor missing somehwere.

[cdg66]

Schematic.PDF
here is the schematic. I know that D- and D+ are swap on the stm32 but they have been patched on the board so that's not the problem

[pigrew]

I usually use SMD 8MHz crystals specified with about 10pF load capacitance. With them, I'd use a pair of 16 or 18 pF caps.

I'm not saying it won't work with your crystal, but verifying the oscillator is what I'd suggest checking. Looking at the RCC status registers can tell you if the oscillator is oscillating, but won't tell you if the frequency is correct.

[cdg66]

the HSEON bit is on in the RCC_CR register

[pigrew]

RCC_CR.HSEON is what you set to request the HSE to be enabled. You'd need to look at RCC_CR.HSERDY, which should be set to 1 if the hardware thinks the HSE is stable.

[cdg66]

RCC_CR.HSERDY is set to 1

[cdg66]

should i call __HAL_RCC_USB_CLK_ENABLE() before the init as its stated in

void dcd_init (uint8_t rhport)
{
/* Clocks should already be enabled /
/ Use __HAL_RCC_USB_CLK_ENABLE(); to enable the clocks before calling this function */

/* The RM mentions to use a special ordering of PDWN and FRES, but this isn't done in HAL.

• Here, the RM is followed. */

[cdg66]

can you tell me for witch specific mcu in the f3 series you acctually tested the stack? I could try to see any difference

[hathach]

https://github.com/hathach/tinyusb/blob/master/docs/boards.md

[cdg66]

I think i found my problem in https://www.st.com/content/ccc/resource/technical/document/application_note/group0/0b/10/63/76/87/7a/47/4b/DM00296349/files/DM00296349.pdf/jcr:content/translations/en.DM00296349.pdf the f2 fammily need a 1.5k pullup to D+ witch i didnt added and were the discovery board do have I'll add one monday and keep you up to date thx for your help.

[cdg66]

After a quick test it confirm that i need a pull up resistor to 3.3v in order to tell the host that im a full speed device otherwise it think that there is nothing pluged in(F3 Don't have it integrated). Thank alot for your help and your quick response.

[pigrew]

Good to hear you figured it out. I had forgotten about the pull-ups.